Enhancement of CO Insertion into a Pd−C Bond in ... - ACS Publications

A series of Pd-containing heterodinuclear methyl complexes, (dppe)(CH3)Pd-MLn (MLn. ) MoCp(CO)3, WCp(CO)3, Co(CO)4; dppe ) 1 ...
0 downloads 0 Views 254KB Size
Organometallics 2001, 20, 2065-2075

2065

Enhancement of CO Insertion into a Pd-C Bond in a Pd-Co Heterodinuclear Complex Atsushi Fukuoka,†,‡ Sumiko Fukagawa,† Masafumi Hirano,† Nobuaki Koga,§ and Sanshiro Komiya*,† Department of Applied Chemistry, Tokyo University of Agriculture and Technology, 2-24-16 Nakacho, Koganei, Tokyo 184-8588, Japan, and Graduate School of Human Informatics, Nagoya University, Nagoya 464-8601, Japan Received December 5, 2000

A series of Pd-containing heterodinuclear methyl complexes, (dppe)(CH3)Pd-MLn (MLn ) MoCp(CO)3, WCp(CO)3, Co(CO)4; dppe ) 1,2-bis(diphenylphosphino)ethane), have been prepared by the metathetical reactions of Pd(CH3)(NO3)(dppe) with Na+[MLn]-, and the complexes were characterized by spectroscopic methods and/or X-ray structure analysis. Related Pt-containing dinuclear complexes were similarly prepared and characterized. The rate of CO insertion into a Pd-CH3 or Pt-CH3 bond was investigated using these complexes. The Pd-Co complex (dppe)(CH3)Pd-Co(CO)4 shows a high activity of CO insertion, giving (dppe)(CH3CO)Pd-Co(CO)4, and the initial rate is ca. 80 times higher than those of the analogous complex Pd(CH3)Cl(dppe). Whereas slow insertion of CO is observed in the PtCo complex (dppe)(CH3)Pt-Co(CO)4, no reaction takes place for Pt(CH3)Cl(dppe). It is revealed by using 13CO that CO in the Co(CO)4 fragment preferentially inserts into the PdCH3 bond over CO in the gas phase. The carbonyl insertion reaction in the Pd-Co heterodinuclear complex has been theoretically studied using B3LYP hybrid density functional theory to clarify its reaction mechanism and the electronic factors controlling the reaction. The calculations for a model complex, (H2PCH2CH2PH2)Pd(CH3)-Co(CO)4, have shown that the most favorable reaction path consists of methyl migration from the Pd to the Co atom, carbonyl insertion reaction on the Co atom, CO coordination to the Co atom, and acetyl migration from the Co atom to the Pd atom, which is in accord with the experimental results. The electron donation from the Pd d orbital to the bridging CO π* orbitals plays an important role in stabilizing the intermediates and transition states. Introduction In transition-metal-catalyzed C-C bond formation reactions,1 Pd complexes play an important role in COrelated reactions such as single and double carbonylation of aryl halide2 and copolymerization of CO with alkenes.3 In these catalytic carbonylation reactions, CO insertion into the Pd-alkyl or Pd-aryl bonds is a key elemental reaction to form C-C bonds, and mono* To whom correspondence should be addressed. E-mail: komiya@ cc.tuat.ac.jp. † Tokyo University of Agriculture and Technology. ‡ Present address: Catalysis Research Center, Hokkaido University, Sapporo 060-0811, Japan. § Nagoya University. (1) (a) Collman, J. P.; Hegedus, L. S.; Norton, J. R.; Finke, R. G. Principles and Applications of Organnotransition Metal Chemistry; University Science Books: Mill Valley, CA, 1987. (b) Yamamoto, A. Organotransition Metal Chemistry; Wiley-Interscience: New York, 1986. (c) Crabtree, R. H. The Organometallic Chemistry of the Transition Metals; Wiley-Interscience: New York, 1994. (d) Parshall, G. W.; Ittel, S. D. Homogeneous Catalysis; 2nd ed.; Wiley-Interscience: New York, 1992. (e) Cornils, B., Herrmann, W. A., Eds. Applied Organometallic Catalysis with Organometallic Compounds; VCH: Weinheim, Germany, 1996. (2) (a) Shoenberg, A.; Bartoletti, I.; Heck, R. F. J. Org. Chem. 1974, 39, 3318-3326. (b) Hidai, M.; Hikita, T.; Wada, Y.; Fujikura, Y.; Uchida, Y. Bull. Chem. Soc. Jpn. 1975, 48, 2075-2077. (c) Ozawa, F.; Soyama, H.; Yamamoto, T.; Yamamoto, A. Tetrahedron Lett. 1982, 23, 3383-3386. (d) Kobayashi, T.; Tanaka, M. J. Organomet. Chem. 1982, 233, C64-C66. (e) Yamamoto, A. Bull. Chem. Soc. Jpn. 1995, 68, 433446.

metallic Pd complexes are widely used with fine-tuning of coordinated ligands. Moreover, it has recently been reported that Pd-containing bimetallic catalysts show unique catalytic activities in carbonylation reactions, which is not observed in the monometallic Pd catalysis. Hidai et al. described that Pd-Co catalysts exhibited high activity in the carbonylation of aryl iodide and that the Pd-Co heterodinuclear acyl complex trans-((CH3)3P)2(C6H5C(O))Pd-Co(CO)4 was formed from the reaction of Pd(C6H5)(OTf)(P(CH3)3)2 with K+[Co(CO)4]-.4 Braunstein et al. reported the copolymerization of CO with norbornene on a Pd-Fe dinuclear complex.5 Heterodinuclear complexes and heterobimetallic cluster complexes have attracted much attention, since they may show a synergistic effect of different metal centers on the reactivity of organic ligands such as alkyl and aryl groups.6 Many papers have reported the synergistic effect of two different metals in catalytic and stoichiometric transformations on the dinuclear or cluster (3) (a) Sen, A.; Lai, T.-W. Z. J. Am. Chem. Soc. 1982, 104, 35203522. (b) Sen, A. Acc. Chem. Res. 1993, 26, 303-310. (c) Drent, E. Eur. Pat. Appl. 121,965, 1984. (d) Drent, E.; van Broekhoven, J. A. M.; Budzulaar, P. H. M. In ref 3e, p 333. (e) Nozaki, K.; Sato, N.; Takaya, H. J. Am. Chem. Soc. 1995, 117, 9911-9912. (4) Misumi, Y.; Ishii, Y.; Hidai, M. J. Chem. Soc., Dalton Trans. 1995, 3489-3496. (5) Braunstein, P.; Knorr, M.; Sta¨hrfekdt, T. J. Chem. Soc., Chem. Commun. 1994, 1913-1914.

10.1021/om001037v CCC: $20.00 © 2001 American Chemical Society Publication on Web 04/19/2001

2066

Organometallics, Vol. 20, No. 10, 2001

complexes. However, there are fewer examples that elucidate the origin of synergism at the molecular level. In our study of the synthesis and reactivity of Ptcontaining heterodinuclear complexes L2RPt-MLn (L2 ) 1,5-cyclooctadiene (cod), dppe; R ) alkyl, aryl; MLn ) MoCp(CO)3, WCp(CO)3, Mn(CO)5, FeCp(CO)2, Co(CO)4), we have demonstrated that reductive elimination and β-H elimination are greatly enhanced on the Pt-M heterodinuclear complexes in comparison to the mononuclear Pt complex.7 In the present work, we extended our research to the chemistry of Pd-containing heterodinuclear complexes, since the Pd-M complexes are expected to show higher catalytic activity than the corresponding Pt-M complexes. We have synthesized new Pd-M heterodinuclear complexes with a CH3 group, and the CO insertion into the Pd-CH3 bonds on the Pd-M complexes has been studied experimentally and theoretically. The CO insertion reaction on the Pd-Co complex is markedly enhanced owing to the heterodinuclear structure, and the reaction mechanism is elucidated by the theoretical calculations. A part of this work was given in a preliminary report.7f Results and Discussion Synthesis and Characterization of Heterodinuclear Complexes. Metathetical reactions of Pd(CH3)(NO3)(dppe), prepared in situ from Pd(CH3)Cl(dppe) (1) and AgNO3, with a slight excess of Na+[MLn]in THF at -30 °C under N2 gave the new dinuclear complexes (dppe)(CH3)Pd-MLn (MLn ) MoCp(CO)3 (2), WCp(CO)3 (3), Co(CO)4 (4)) (eq 1). When spectroscopic

and analytical data of 2-4 are compared with those of the related Pt analogues (dppe)(CH3)Pt-MLn,7c it is established that 2-4 have similar dinuclear structures with a methyl group on Pd, as shown in eq 1. Elemental analyses gave satisfactory results for 2 and 4. The molar electric conductivities of 2-4 were significantly low, showing that they are not ionic but neutral complexes. Selected NMR and IR data of 2-4 are listed in the Experimental Section. Typically, the 1H NMR spectrum (6) (a) Braunstein, P., Oro, L. A., Raithby, P. R., Eds. Metal Clusters in Chemistry; Wiley-VCH: Weinheim, Germany, 1999. (b) Adams, R. D., Cotton, F. A., Eds. Catalysis by Di- and Polynuclear Metal Cluster Complexes; Wiley-VCH: New York, 1998. (c) Chetcuti, M. J. In Comprehensive Organometallic Chemistry II; Adams, R. D., Ed.; Pergamon Press: Oxford, U.K., 1995; Vol. 10. (d) Shriver, D. F., Kaesz, H., Adams, R. D., Eds. The Chemistry of Metal Cluster Complexes; VCH: New York, 1990. (e) Stephan, D. W. Coord. Chem. Rev. 1989, 95, 41-107. (7) (a) Komiya, S.; Endo, I. Chem. Lett. 1988, 1709-1712. (b) Miki, K.; Kasai, N.; Endo, I.; Komiya, S. Bull. Chem. Soc. Jpn. 1989, 62, 4033-4035. (c) Fukuoka, A.; Sadashima, T.; Sugiura, T.; Wu, X.; Mizuho, Y.; Komiya, S. J. Organomet. Chem. 1994, 473, 139-147. (d) Fukuoka, A.; Sadashima, T.; Endo, I.; Ohashi, N.; Kambara, Y.; Sugiura, T.; Miki, K.; Kasai, N.; Komiya, S. Organometallics 1994, 13, 4033-4044. (e) Fukuoka, A.; Sugiura, T.; Yasuda, T.; Taguchi, T.; Hirano, M.; Komiya, S. Chem. Lett. 1997, 329-330. (f) Fukuoka, A.; Fukagawa, S.; Hirano, M.; Komiya, S. Chem. Lett. 1997, 377-378. (g) Yasuda, T.; Fukuoka, A.; Hirano, M.; Komiya, S. Chem. Lett. 1998, 29-30. (h) Komiya, S.; Muroi, S.; Furuya, M.; Hirano, M. J. Am. Chem. Soc. 2000, 122, 170-171.

Fukuoka et al. Table 1. Selected Bond Distances (Å) and Angles (deg) for 5 Pt-Co Pt-P(2) Pt-C(2) Pt-C(4) Co-C(2) Co-C(4) Co-Pt-P(1) P(1)-Pt-P(2) Co-C(2)-O(1) Co-C(4)-O(3)

2.676(2) 2.235(4) 2.66(2) 2.64(2) 1.79(2) 1.81(2) 100.4(1) 85.5(2) 174(1) 170(1)

Pt-P(1) Pt-C(1) Pt-C(3) Pt-C(5) Co-C(3) Co-C(5) Co-Pt-C(1) P(2)-Pt-C(1) Co-C(3)-O(2) Co-C(5)-O(4)

2.313(5) 2.12(2) 4.24(2) 3.73(2) 1.79(2) 1.81(2) 85.7(4) 88.3(5) 177(2) 177(2)

of (dppe)(CH3)Pd-Co(CO)4 (4) gave a doublet of doublets at δ 1.21 due to the methyl group coupled to two phosphorus nuclei of dppe (3JPH ) 7.7, 5.0 Hz). In the 31P{1H} NMR spectrum of 4, the dppe resonances were observed at δ 37.2 and 49.4 as two doublets (2JPP ) 26 Hz). These results show that the two P nuclei of dppe are not magnetically equivalent, implying a squareplanar geometry around Pd. The IR spectrum of 4 gave four ν(CO) bands at 2020, 1953, 1913, and 1885 cm-1, which are similar to those for -I valent Na+[Co(CO)4](2002 cm-1). We infer that the oxidation states of the metals are close to Pd(II) and Co(-I), although the formal oxidation states are Pd(I) and Co(0) in the bimetallic structure with a direct Pd-Co bond. The presence of the Pd-Co bond has been confirmed by X-ray crystallography, as described below. The molecular structures of the Pd-Co complex 48 and its Pt-Co analogue (dppe)(CH3)Pt-Co(CO)47c (5) were determined by single-crystal X-ray structure analyses. ORTEP drawings of 4 and 5 are depicted in Figure 1, and selected bond distances and angles for 5 are listed in Table 1. The Pd-Co distance for 4 is 2.682(8) Å, which is slightly shorter than that for ((CH3)3P)2(C6H5C(O))Pd-Co(CO)4 (2.7856(7) Å).4 This result unequivocally demonstrates the presence of a Pd-Co single bond in 4. The molecular structure of the corresponding PtCo complex 5 is basically isomorphous with 4, and the axial C(4)-O(3) and C(2)-O(1) carbonyls are slightly bent, indicating a weak semibridging interaction with Pt: Pt-C(4) ) 2.64(2) Å and Co-C(4)-O(3) ) 170(1)°; Pt-C(2) ) 2.66(2) Å and Co-C(2)-O(1) ) 174(1)°. CO Insertion Reaction into the Pd-CH3 Bond. The reactivity of the new heterodinuclear complexes toward CO insertion was investigated. Treatment of the Pd-Co complex 4 with atmospheric CO in benzene at room temperature led to the formation of an acetyl complex, (dppe)(CH3C(O))Pd-Co(CO)4 (6), in 16 h (eq 2).

Complex 6 was isolated as orange needles in 85% yield by recrystallization from a mixture of benzene and hexane. Under the same reaction conditions, the Pd(8) Due to the small size and low quality of the single crystals for 4, the diffractometer could not collect enough reflection data (reflection/ parameter ratio 3.0) even though the R value and GOF (R ) 0.073, GOF ) 1.87) were reasonable. Therefore, we have avoided detailed discussion concerning the bond distances and angles, except for the Pd-Co bond and the overall structure of 4.

CO Insertion into a Pd-C Bond

Organometallics, Vol. 20, No. 10, 2001 2067

Figure 1. ORTEP drawings of (dppe)(CH3)Pd-Co(CO)4 (4) and (dppe)(CH3)Pt-Co(CO)4 (5).

Figure 2. Time-yield curves of CO insertion.

Mo complex 2 gave the CO insertion product (dppe)(CH3C(O))Pd-MoCp(CO)3 (22%) and the reductive elimination product Mo(CH3)Cp(CO)3 (27%). In contrast, no CO insertion took place for the Pd-W complex 3, but W(CH3)Cp(CO)3 was quantitatively obtained from reductive elimination. For the Pt-Co complex 5, the CO insertion proceeded at a slow rate to give (dppe)(CH3C(O))Pt-Co(CO)4 (60%) in 10 days. From these results, it is shown that the presence of a Co(CO)4 moiety provides good activity and selectivity of CO insertion to form the heterodinuclear acetyl complexes. The reaction rates of CO insertion for several PdCH3 and Pt-CH3 complexes were monitored by NMR spectroscopy, employing samples of the complexes with concentrations of ca. 0.02 M in CD2Cl2 at 24 ( 1 °C and 1 atm (ca. 9-fold excess CO). Figure 2 represents timeyield curves of CO insertion to give the corresponding acetyl complexes. The Pd-Co complex 4 gave a signifi-

cantly higher activity than the mononuclear Pd complex 1 and the Pt-Co complex 5 with methyl and dppe ligands; the initial rate of CO insertion for 4 was approximately 80 times faster than that of 1. No CO insertion occurred for PtMeCl(dppe) under the same conditions. Hence, it is clearly shown that the CO insertion is enhanced by the Co(CO)4 group over the Cl ligand for the methyl complexes of Pd and Pt with dppe ligands. It is well-known that cationic Pd(II) complexes show a high activity of CO insertion. With this mechanism for the formation of 6 from 4 and CO, heterolytic cleavage of the Pd-Co bond may be possible to generate [Pd(CH3)(dppe)]+[Co(CO)4]-, and subsequent CO insertion on the Pd cation9 and recombination of Pd-Co bond would provide 6. In this case, addition of [Co(CO)4]would suppress the CO insertion. However, when 1 equiv of Na+[Co(CO)4]- or [PPN]+[Co(CO)4]- (PPN ) (Ph3P)2N) was added to the reaction of 4 with CO, the time-yield curves were almost the same as that of 4 itself, as shown in Figure 2. Thus, the intermediacy of the cationic Pd(II) species is not likely to occur under our reaction conditions.10 Accordingly, it is conceivable that the Pd-Co bond is retained in the CO insertion reaction. Isotopic Labeling Experiment for CO Insertion. In the mechanism of the CO insertion on 4 to form 6, the acetyl CO results from either gaseous CO or Co(CO)4. Therefore, to reveal the origin of acetyl CO, isotopic labeling experiments were performed. When 4 was treated with 13CO (27-fold excess, 99% 13C-enriched) in benzene in a Schlenk tube at room temperature, a 13C-enriched acetyl Pd-Co complex (6*) was isolated in 82% yield in 6 h (eq 3). The 13C content of the acetyl (9) (a) Dekker, G. P. C. M.; Elsevier, C. J.; Vrieze, K.; van Leeuwen, P. W. N. M. Organometallics 1992, 11, 1598-1603. (b) Kayaki, Y.; Kawataka, F.; Shimizu, I.; Yamamoto, A. Chem. Lett. 1994, 21712174. (c) Kayaki, Y.; Shimizu, I.; Yamamoto, A. Chem. Lett. 1995, 1089-1090. (d) Markies, B. A.; Kruis, D.; Rietvelt, M. H. P.; Verkerk, K. A. N.; Boersma, J.; Kooijman, H.; Lakin, M. T.; Spek, A. L.; van Koten, G. J. Am. Chem. Soc. 1995, 117, 5263-5274.

2068

Organometallics, Vol. 20, No. 10, 2001

Figure 3. IR (a) and

31P{1H}

NMR spectra (b) of

Fukuoka et al.

13C-enriched

CO carbon was calculated by IR and 31P{1H} NMR spectroscopy. By comparison of the absorbance of acetyl groups (ν(12CO) 1674 cm-1, ν(13CO) 1639 cm-1) in the

IR spectrum of 6*, the 12C/13C ratio was estimated to be 2.5 (Figure 3a). In the 13C{1H} NMR spectrum of 6*, the resonance due to C(O)CH3 at δ 235.0 showed 31P13C couplings (2J PC ) 12 and 115 Hz) with the phosphorus nuclei of dppe. On the other hand, in the 31P{1H} NMR spectrum of 6* (Figure 3b), an AB pattern at δ 24.2 and 25.9 was observed as major species for the phosphorus nuclei trans to the acetyl and cobalt moieties, respectively, in which one doublet of doublets due to the 13C-31P (2JCP ) 115 Hz) and the 31P-31P (2JPP ) 44 Hz) couplings was also observed at 24.2 ppm as satellites. From the integral value of this peak, the 12C/ 13C ratio was estimated to be 2.0. Both the IR and 31P{1H} NMR data indicate that most of the acetyl carbonyl is 12CO from Co(CO)4. If 12CO and 13CO were statistically scrambled in the reaction mixture, the 12C/ 13C ratio would be 0.15 (4 12CO from Co(CO) and 27 4 13CO from the gas phase). It is of great interest to note that the CO ligand in Co(CO)4 preferentially inserts into the Pd-CH3 bond even in the presence of 7-fold excess 13CO in the gas phase. We infer that the minor 13C(O)(10) Other mechanisms for the CO insertion on 4 are as follows. (1) dppe becomes monodentate and the semibridging CO migrates to Pd to result in CO insertion. To check this mechanism, we added 1 equiv of dppe or P(CH3)3 to 4, but ionization took place to give the ionic complexes [(dppe)(CH3)Pd(µ-dppe)Pd(CH3)(dppe)]+[Co(CO)4]- and [Pd(CH3)(P(CH3)3)(dppe)]+[Co(CO)4]-, respectively. Thus, it was difficult to prove this mechanism. However, we infer that the ionization to cationic Pd(II) species is a minor process, because (dppe)(CH313C(O))Pd-Co(CO)4 was a minor product in the reaction of 4 with 13CO (vide infra). [Pd(CH3)(P(CH3)3)(dppe)]+[Co(CO)4]- showed no activity of CO insertion under the conditions of Figure 2. (2) CO enhances the reductive elimination of Co(CH3)(CO)4 from 4 to give “Pd(dppe)” species, and CO insertion and coordination of CO provide Co(C(O)CH3)(CO)4. Oxidative addition of the acetyl Co complex to “Pd(dppe)” gives (dppe)(CH3C(O))Pd-Co(CO)4. This is still a possible way to explain the results of the 13CO labeling experiment.

(dppe)(CH3C(O))Pd-Co(CO)4 (6*).

CH3 species is formed by the direct attack of 13CO at Pd or scrambling of gaseous 13CO with the 12CO ligand on Co(CO)4, but the isotopic labeling experiments show that they are slower processes. Theoretical Calculation for the Carbonyl Insertion Reaction into Pd-CH3 Bond. To elucidate the reaction mechanism of the carbonyl insertion reaction into the Pd-CH3 bond to form the acetyl group on the Pd atom, we performed B3LYP density functional calculations for the reaction (eq 4) of a model reactant, (H2PCH2CH2PH2)Pd(CH3)-Co(CO)4 (7), in which the phenyl groups of the bis-phosphine ligand are replaced by hydrogen atoms.

(a) Model Reactant (H2PCH2CH2PH2)Pd(CH3)Co(CO)4. The optimized structure of the model reactant (H2PCH2CH2PH2)(CH3)Pd-Co(CO)4 (7), shown in Figure 4, is qualitatively in agreement with the X-ray structure of (dppe)(CH3)Pd-Co(CO)4 (4).7f For instance, the Pd-Co bond length in the experimental structure is 2.682(8) Å, whereas the calculated length is 2.643 Å. To analyze the electronic structure of 7, we localized the Kohn-Sham orbitals using the Boys algorithm.11 The two orbitals representing the Pd-Co bond and the interaction between the Pd atom and the semibridging carbonyl groups are shown in Figure 5.12 One can see in Figure 5a that there is a Pd d-Co d σ bond and that the bonding electrons delocalize into the π* orbitals of the semibridging carbonyl ligands. As shown in Figure 5b, similar electron back-donation from the occupied Pd d orbitals to the π* orbitals takes place as well. Although one may ascribe the semibridging structure to this back-donation from the Pd atom, this is not the sole origin of semibridging. The structure of CoH(CO)4 (11) (a) Boys, S. F. Rev. Mod. Phys. 1960, 32, 296-299. (b) Foster, J. M.; Boys, S. F. Rev. Mod. Phys. 1960, 32, 300-302. (12) The localized orbitals obtained from the RHF orbitals are qualitatively the same as those based on the Kohn-Sham orbitals.

CO Insertion into a Pd-C Bond

Organometallics, Vol. 20, No. 10, 2001 2069 Chart 1

Scheme 1

Figure 4. Optimized structure of (H2PCH2CH2PH2)(CH3)Pd-Co(CO)4 (7), with bond distances in Å. The experimental bond distances are in parentheses.

Figure 5. Localized orbitals representing (a) the Pd-Co bond and (b) donation from the occupied Pd d orbital to CO π* orbitals.

has been studied theoretically13 to show that, although it is not an equilibrium structure, CoH(CO)4 with an equatorial hydride (C2v symmetry) displays bending of CO ligands toward the H ligand.14 The calculations at the present level gave an H-Co-COax angle of 73.0° in CoH(CO)4, as shown in Chart 1 , and also gave C-CoCOax angles of 78.5 and 78.5-78.7° in Co(CH3)(CO)4 and Co(CH(CH3)2)(CO)4, respectively. These results show that CO bending is normal for tetracarbonylcobalt complexes with an equatorial σ bond and that back(13) Torrent, M.; Sola, M.; Frenking, G. Chem. Rev. 2000, 439-493. (14) (a) Elian, M.; Hoffmann, R. Inorg. Chem. 1975, 14, 1058-1076. (b) Versluis, L.; Ziegler, T.; Baerends, E. J.; Ravenek, W. J. Am. Chem. Soc. 1989, 111, 2018-2025.

donation in 7 plays a secondary role to reduce the angles to 67-68°. However, electron-back-donation plays an energetically important role. The C2v structure of CoH(CO)4 with an equatorial Co-H σ bond is not an equilibrium structure but a transition state connecting the two C3v equilibrium structures, which are 7.9 kcal/mol more stable than the C2v structure. This indicates that the structure of the Co moiety with an equatorial Pd-Co σ bond is an equilibrium structure due to back-donation in 7. Starting from 7, there are three possible reaction paths shown in Scheme 1. In path 1 the methyl group on the Pd atom migrates to the Co atom so that the carbonyl insertion takes place on the Co atom. Finally the acetyl group migrates from the Co atom to the Pd atom to lead to the product. In path 2 one of the carbonyl groups on the Co atom migrates to the Pd atom and then the carbonyl insertion takes place on the Pd atom. In path 3 the CC bond is directly formed between the two groups on the two metals. We have theoretically studied these three reaction paths. We performed theoretical calculations at the B3LYP density functional level using set I for structure determinations and the basis set II for final energy calculations. In the following discussions we will refer to the energies calculated at the B3LYP/ II//B3LYP/I level, unless otherwise mentioned. (b) Reaction Starting with the Methyl Migration from Pd to Co: Path 1.

2070

Organometallics, Vol. 20, No. 10, 2001

Fukuoka et al.

Figure 7. Localized orbitals for electron back-donation from the occupied Pd d orbital to the CO π* orbitals in 8b.

Figure 6. Optimized structures of the transition state and product for methyl migration in path 1. Energies are relative to 7 in kcal/mol at the B3LYP/II(B3LYP/I) level. Chart 2

The first step of path 1 is methyl migration from the Pd atom to the Co atom. The optimized structures for this step are shown in Figure 6. The methyl migration passes through the transition state, TS(7f8a), with an activation energy of 15.4 kcal/mol. At TS(7f8a) the methyl group bridges the two metal atoms, as expected for migration reactions, with Pd-CH3 and Co-CH3 distances of 2.339 and 2.365 Å, respectively. Also, one of the CO groups (CO1) bridges the two metal atoms. The product of this step is intermediate 8a, in which the migrating methyl group occupies the axial position of the Co moiety with a deformed trigonal-bipyramidal (TBP) structure, a structure favorable for d8 fivecoordinate complexes such as CoH(CO)4 mentioned above. In 8a CO1 continues to bridge the metal atoms. There is another isomeric structure with the two CO bridges, 8b, shown in Figure 6. The isomerization from 8a to 8b, which includes the rotation of the TBP Co moiety and results in bond formation between the Pd atom and the CO3 carbon atom, is 0.5 kcal/mol endothermic, and its activation energy is only 0.7 kcal/mol. These small values indicate that the potential energy surface in this region is quite flat and that fluxional behavior is expected. For 8 with CO bridge(s) we have two possible ways of drawing chemical bonds, as shown in Chart 2. (i) The electron donation from Pd d orbital to the π* orbital of the bridging CO groups is responsible for the bonding interaction between two fragments in addition to the bonds between the metal atoms, shown by broken lines. (ii) On the other hand, the bridging CO groups are ketonic. The Pd and Co atoms in (i) have d10 and d8

formal electron configurations, respectively, whereas those in (ii) have d8 and d6 configurations, respectively. To know which is a better model, we again calculated localized orbitals to find that the first model properly represents the interaction between the two fragments. For instance, the localized orbitals shown in Figure 7 demonstrates the electron donation from the occupied Pd d orbital to the π* orbitals of the bridging CO groups in 8b, whereas we did not obtain a localized orbital representing a Pd-Co σ bond. d10 ML2 complexes with an L-M-L angle of about 90° like the Pd fragment in 8a and 8b have an unstable occupied d orbital which is known to activate even CH bonds. The electron backdonation to the π* orbitals stabilizes such an unstable Pd fragment. This is an energetically important factor of the reaction, and stabilization due to this electron back-donation is observed in other intermediates and TSs in the present path. In the second step of path 1 the bond between the methyl group on the Co atom and one of the CO groups is formed. The methyl group in 8a and 8b can react with bridging or terminal CO groups. The structures for three reaction paths, one from 8a and two from 8b, are shown in Figure 8. The reaction with terminal CO2 in 8b, which gives 9b, is just a migratory carbonyl insertion into the Co-CH3 bond. During the course of this reaction the carbonyl group does not insert into the bond, but the methyl group migrates to the CO2 group. This reaction passes through three-centered TS(8bf9b) with an activation energy of 8.9 kcal/mol. While the H3C-CO2 bond distance is 1.954 Å, which is 1.25 times longer than that of 1.559 Å in the product, the Co-CH3 bond of 2.248 Å is only 1.081 times longer compared with that of 2.080 Å in 8b. While the H3C-CO2 bond is partially formed, the Co-CH3 bond is hardly stretched. Asynchronous bond exchange like this has been theoretically found in the TSs for various CO migratory insertions.15 The local structure of the Co fragment of this TS is square pyramidal with the apical CO4 group, and the bond exchange takes place in the basal plane. The reaction with another terminal CO group, CO4, is expected to be similar, and therefore we did not investigate this reaction. The reaction of the methyl group with the bridging carbonyl group, CO1, in 8b was shown to take place with an activation energy of 8.3 kcal/mol, only 0.6 kcal/mol smaller than that for the reaction with the terminal CO2 group. Also, the H3C-CO1 bond distance of 1.925 Å and the Co-CH3 bond distance of 2.238 Å in the transition state, TS(8bf9a), with a square-pyramidal structure of the Co moiety, are similar to those in TS(8bf9b). At (15) Koga, N.; Morokuma, K. Chem. Rev. 1991, 91, 823-842.

CO Insertion into a Pd-C Bond

Organometallics, Vol. 20, No. 10, 2001 2071

Figure 8. Optimized structures for CO insertion reactions from 8a and 8b. The bond distances and angles are given in Å and deg, respectively. The energies are relative to 7 at the B3LYP/II(B3LYP/I) level.

this TS the Pd-CO1 bond is stretched by 0.22 Å, and the CoCO1 fragment is more linear compared with that in 8b. These results show that the CO1 group at TS(8bf9a) is similar to the terminal carbonyl in character. In a comparison between 8a and 8b, one notices that the second CO bridging easily breaks. Accordingly, it is not surprising that these two reactions are energetically similar. During the course of the reaction from 8b to 9a the Co fragment rotates so that CO2 bridges the metal atoms in 9a, suggesting that the acetyl group is not sufficiently electron-accepting to form bridges. 9a and 9b are enantiomeric, except for the conformation of the ethylene bridge in the bis-phosphine ligand. Therefore, these two intermediates are isoenegetic. In 9a and 9b the agostic interaction between the β-C-H bond of the acetyl ligand and the Co atom takes place as shown by the short Co-Hβ distance of 2.383 and 2.379 Å in 9a and 9b, respectively. We can consider three reaction pathways from 8a as well. However, we found only one transition state structure, TS(8af9b). In several trials for finding other transition states we obtained only TS(8bf9a) and TS(8bf9b), showing that at the TSs CO bridging is newly formed to stabilize the reaction system. On the other hand, in the reaction 8af9b, CO3 bridges the metal atoms after passing through TS(8af9b). The activation energies for all these CO insertion reactions are similar, 8.3-10.0 kcal/mol, and thus we can say at the present level of calculations that all the insertion reactions are possible. Since these reactions are on the single metal atom, the profile of the reactions is expected to be similar to that for the reaction of mononuclear complexes. As a matter of fact, the calculations for the carbonyl insertion reaction of Co(CH3)(CO)4 at the same level of calculations showed that the mononuclear reaction requires an activation energy of 10.8 kcal/mol, to show that the Pd fragment has a small electronic effect on the activation energy. The TS for the reaction of Co(CH3)(CO)4 has a square-pyramidal framework, and the product has an agostic acyl ligand with a

Co-Hb distance of 2.272 Å.16 They are the same geometrical features as found in the TSs of the Pd-Co complex. The bond formation in the migratory CO insertion takes place on the Co atom, and thus the resultant acetyl group has to migrate to the Pd atom. In addition, the extra CO group has to coordinate to the Co atom, to lead to the final product. Therefore, there are two possibilities for reaction paths. In one path (path a), first CO coordinates to the Co atom followed by acetyl migration, and in another path (path b) acetyl migration prior to CO coordination takes place. Concerned with path a, as shown in Figure 9, we found the transition state TS(10bf11) for the acetyl migration. The reactant corresponding to this TS is 10b, in which the Co moiety has a deformed TBP structure and one of the CO groups bridges the metal atoms. At the B3LYP/I level we found another equilibrium structure, 10a, with the two CO bridges. This is considered to be the intermediate formed by CO coordination to 9, and it is 18.7 kcal/mol more stable than 9a + CO. CO groups in 10a are numbered by assuming that the fifth CO group coordinates to the Co atom between CO2 and the acetyl groups of 9a. There is a transition state between 10a and 10b, TS(10af10b), at the B3LYP/I level. However, at the B3LYP/II//B3LYP/I level TS(10af10b) is more stable than 10a, indicating that 10a does not exist as an equilibrium structure and that CO coordination to 9 directly results in the structure with a single CO bridge, 10b. Also, the small energy difference between 10a and 10b suggests that, similar to the case of 8a and 8b, the potential energy surface is flat and the structure is fluxional. During the process from 10b to TS(10bf11) the Co moiety rotates around the Co-CO3 bond so that the migrating acetyl group bridges the metal atoms. While the reaction from 10b is 7.3 kcal/ (16) In the case of Co(C(O)CH3)(CO)4 the structure with a Co-O interaction is more stable than that with an agostic interaction. This is true in the Pd-Co complex. The structure with the Co-O interaction is 4 kcal/mol more stable than 9a. However, the transition state for the acetyl migration is not connected to it, and therefore we omitted it in discussions.

2072

Organometallics, Vol. 20, No. 10, 2001

Fukuoka et al.

Figure 10. Energy profile for the favorable path for reaction 4 calculated at the B3LYP/II//B3LYP/I level. Enthalpies in italics are those at the B3LYP/I//B3LYP/I level, and enthalpies in parentheses are those calculated at 298.15 K.

Figure 9. Optimized structures for acetyl migration from the Co atom to the Pd atom. In path a the migration follows the CO coordination, and in path b the acetyl migration first takes place in 9a. The energies are relative to 7 + CO for the reactions from 10a and relative to 7 for the reactions from 9a at the B3LYP/II(B3LYP/I) level.

mol exothermic, the activation energy for the acetyl group migration from 10b was calculated to be 5.9 kcal/ mol. In the final product 11 the two carbonyl groups semibridge the metal atoms as in the reactant, 7. This acetyl group migration from the Co atom to the Pd atom is analogous to the reverse reaction, 8af7, of the methyl migration. The distances of the COCH3 group from the metal atoms in TS(10bf11) are shorter than those of the CH3 group in TS(7f8a), suggesting that the CO π orbitals of the acetyl group interact with the metal d orbitals to stabilize the transition state. The smaller activation energy for 11f10b than that for 7f8a is consistent with this. If the acetyl group migration takes place prior to the CO coordination, the reaction passes through TS(9af12), shown in Figure 9. While the acetyl group in TS(9af12) bridges the two metal atoms as in TS(10bf11), the reaction from 9a to 12 requires 3 kcal/ mol greater activation energy, being less favorable than the reaction from 10b to 11. Different from the exothermic reaction 10bf11, the reaction 9af12 is almost thermoneutral. This is presumably because this reaction gives a coordinatively unsaturated Co fragment. In the last stage of path b CO coordination takes place with an exothermicity of 26 kcal/mol to lead to the final product. (c) Comparison of Path 1 with the Other Paths. We extensively searched transition states for the CO

migration in path 2 from the Co atom to the Pd atom to find only one transition state in which the P2 atom almost dissociates from the Pd atom and CO3 migrates to the vacant site thus generated. The activation energy of this CO migration (21 kcal/mol) is larger than that of the methyl migration of path 1 and thus less favorable. In the present system the five-coordinated Pd fragment is too unstable to exist. Similarly, during the structure determination of (OC)(H2PCH2CH2PH2)(CH3)Pd-Co(CO)4, one of the P atoms dissociates. The activation barrier to the direct C-C bond formation between the CH3 group on the Pd atom and the CO ligand on the Co atom in path 3 was calculated to be 25 kcal/mol, which is the largest. In addition, during the course of the reaction one of the P atoms migrates to the Co atom. It is unlikely that such an elementary step is the part of the present reaction. Consequently, we can conclude that path 1 is the reaction path for the acetyl group formation in the Pd-Co complex. (d) Energy Profile. The profile of potential energy as well as enthalpy calculated at 298.15 K for the favorable reaction path is summarized in Figure 10. As for the CO migratory insertion there are several comparable reaction paths, and the profile for 8bf9a was adopted in Figure 10. The first step of the methyl migration is rate-determining, which requires an activation energy of 15.4 kcal/mol. The CO insertion on the Co atom and the following acetyl migration are easier. The effects of basis set are small for the intramolecular migration and insertion reactions, whereas those for the CO coordination is not negligible. Because of the basis set superposition error the CO binding energy is overestimated using the basis set I. The larger basis set II improves the energetics. Thermal contributions taken into account in enthalpy do not change the profile qualitatively. In the present reaction, the Co fragment is chosen as the site for the C-C bond formation, and the methyl group migration between the two metal centers is a key reaction. The important factor which realizes the methyl migration is electron back-donation from the Pd occupied orbital to the CO π* orbitals as discussed above, which stabilizes the unstable d10 Pd fragment. Accordingly, a similar migration could take place in complexes with late-transition-metal atoms such as Pd. Though of lower reactivity, the same reaction is experimentally known for the Pt-Co analogue as mentioned above. Our B3LYP/I calculations showed that the methyl migration

CO Insertion into a Pd-C Bond

Organometallics, Vol. 20, No. 10, 2001 2073

higher than that for the Pd-Co system and smaller than that for the Pt-Co system. However, the difference is found in the structure of the bis-phosphine fragment in the TS as well as in the product. The P2 atom dissociates, the corresponding Pd-P2 distance at the TS and product being 3.783 and 3.914 Å, respectively. The earlier, more electropositive W atom donates its d electrons into the π* orbitals of CO more than the Co atom.18 Because of this stronger back-donation the carbonyl C-O distances of 1.20 Å in the Pd-W complex are longer than those in 7. Electron-rich CO groups on the W atom cannot stabilize the Pd fragment by accepting electrons from the Pd atom, and therefore, to avoid the unstable structure of the Pd fragment the phosphorus atom dissociates. The interaction between the Pd and W fragments in 14 should be weaker than that between the Pd and Co fragments. Actually the interaction between the two metal fragments in 14 was calculated to be weaker by 8-9 kcal/mol, as shown in eqs 6 and 7, and the

Figure 11. Optimized structures for methyl migration in (H2PCH2CH2PH2)(CH3)Pd-WCp(CO)3, (13). The distances and angles are given in Å and deg, respectively. The energies are relative to 13.

(eq 5) is possible, but the activation barrier (24.3 kcal/ mol) is 9 kcal/mol larger than that of the Pd-Co system

dissociation into Pd and W fragments is possible, different from the Pd-Co system. This result shows that the present reaction depends on the nature of the transition metals and accounts for the experimental observation that W(CH3)Cp(CO)3 was obtained from reductive elimination. Concluding Remarks

in agreement with the experimental observation. The stronger Pt-C bond results in a greater activation energy. (e) Comparison of Methyl Migration with That of the Pd-W Complex. These results suggest that the situation could be different in heterodinuclear complexes with earlier transition metals. The heterodinuclear Pd-W complex similar to the present reactant is experimentally known, in which the methyl group is on the Pd atom and three carbonyl groups are on the W atom.17 As mentioned above, the heterodinuclear Pd-W complex 3 does not conduct CO insertion. Thus, to elucidate the origin of the difference, we studied similar model methyl migration for (H2PCH2CH2PH2)(CH3)PdWCp(CO)3 (13), as shown in Figure 11. The activation energy was calculated to be 18 kcal/mol, not much (17) Komiya, S. Unpublished results. Reaction of Pd(CH3)Cl(cod) with Na+[WCp(CO)3]- also gave (cod)(CH3)Pd-WCp(CO)3, which slowly afforded W(CH3)Cp(CO)3 during the purification process. The Pt analogue (cod)(CH3)Pt-WCp(CO)3 reacted with CO, giving W(CH3)Cp(CO)3 in good yield, but no insertion product was obtained.7d

We have studied the cooperative effects in the PdCo, Pd-Mo, and Pd-W heterodinuclear complexes. The treatment of 4 with 13CO resulted in the preferential insertion of the carbonyl on Co into the Pd-C bond and the incoming CO locates on Co. The B3LYP hybrid density functional calculations suggest that this insertion reaction proceeds via methyl migration from Pd to Co, insertion of a carbonyl into the CH3-Co bond, coordination of incoming carbon monoxide to the vacant site in Co, and migration of the acetyl group from Co to Pd, which is in good agreement with the experimental results. Such facile migration of the organic group depends on the nature of the transition-metal fragment. In the Pd-W system the activation energy for the methyl migration is as much as that in the Pd-Co system. However, the more electron rich W fragment dispenses its electrons to back-donation to the carbonyl (18) (a) Pauling electronegativities for Co and W are 1.8 and 1.7, respectively. Pauling, L. The Nature of the Chemical Bond; Cornell University Press: Ithaca, NY, 1960. (b) Orbital energies (in au) for neutral W and Co atoms are as follows: W, 5d (-0.45), 6s (-0.22); Co, 3d (-0.68), 4s (-0.27). Fraga, S.; Karwowski, J.; Saxena; K. M. S. Handbook of Atomic Data; Elsevier: Amsterdam, 1976.

2074

Organometallics, Vol. 20, No. 10, 2001

groups and cannot stabilize the Pd-W bond, leading to a simple alkyl transfer reaction. These results will provide a fundamental understanding of the synergistic effect of the related bimetallic catalysis. Experimental Section All manipulations were carried out under an atmosphere of nitrogen or argon using standard Schlenk techniques.1b,19 Solvents were dried over and distilled from appropriate drying agents under N2: hexane, benzene, toluene, diethyl ether, and THF from Na/benzophenone ketyl; CH2Cl2 from P2O5. NMR solvents were freeze-pump-thaw-degassed and vacuumtransferred from appropriate drying agents (C6D6 from Na; CDCl3 and CD2Cl2 from P2O5). 1,2-Bis(diphenylphosphino)ethane (dppe) was prepared by the literature method.20 [PPN]+Cl- (PPN ) (Ph3P)2N), CO (99.9%), and 13CO (99%) were used as received. IR spectra were measured on a JASCO FT-IR 5M spectrometer. NMR spectra were obtained on a JEOL FX-200 (1H, 199.5 MHz) or a JEOL LA-300 (1H 300.4 MHz) spectrometer. Chemical shifts were relative to Si(CH3)4 (1H and 13C) and 85% H3PO4 (31P). Elemental analyses were performed with a Perkin-Elmer 2400 series II CHN analyzer. Molar electric conductivity was measured on a TOA CM-7B conductimeter. The starting materials and related complexes were prepared by the literature methods with minor modifications: Pd(CH3)Cl(cod),21 Pt(CH3)Cl(dppe),22 Pt(CH2CH3)Cl(dppe),23 Na+[MoCp(CO)3]-,24 Na+[WCp(CO)3]-,25 Na+[Co(CO)4]-,25 and [PPN]+[Co(CO)4]-.26 Pd(CH3)Cl(dppe) was prepared by the ligand substitution reaction of Pd(CH3)Cl(cod) with an equimolar amount of dppe in benzene at room temperature. The precipitated white powder was washed with hexane, and recrystallization from a mixture of CH2Cl2 and hexane gave white crystals (yield 98%). Synthesis of Heterodinuclear Complexes. A typical procedure for (dppe)(CH3)Pd-Co(CO)4 (4) is given. To a THF solution of Pd(CH3)Cl(dppe) (460.8 mg, 0.830 mmol) was added AgNO3 (141.7 mg, 0.834 mmol). Stirring at room temperature gave a colorless solution with a white precipitate of AgCl. To the filtered solution was added a THF solution of Na+[Co(CO)4](182.4 mg, 0.926 mmol) at -20 °C, and the mixture was stirred for 2 h. The reaction mixture was evaporated to dryness, and the resulting solid was extracted with benzene. Addition of hexane to the concentrated solution of extracts gave red cubes. The crystals were washed with cold hexane and dried under vacuum (369.1 mg, 0.534 mmol). Yield: 64%. Mp: 123 °C dec. Anal. Calcd for C31H27O4P2CoPd: C, 53.90; H, 3.94. Found: C, 53.98; H 3.96. Molar electric conductivity Λ (THF, 24 °C): 0.017 S cm2 mol-1. IR (ν(CO), KBr): 2020, 1953, 1913, 1885 cm-1. 1H NMR (C6D6): δ 1.21 (dd, 3JPH ) 7.7, 5.0 Hz, 3H, CH3), 1.6-1.9 (m, 4H, dppe CH2), 7.0-7.6 (m, 20H, dppe C6H5). 31P{1H} NMR (C6D6): δ 37.2 (d, 2JPP ) 26 Hz), 49.4 (d, 2JPP ) 26 Hz). 13C{1H} NMR (C6D6): δ 9.8 (d, 2JPC ) 106 Hz, CH3), 27.8 (dd, JPC ) 24, 16 Hz, dppe CH2), 29.2 (dd, JPC ) 26, 22 Hz, dppe CH2), 129-134 (m, dppe C6H5), 210 (br s, CO). (19) (a) Komiya, S., Ed. Synthesis of Organometallic Compounds: A Practical Guide; Wiley: New York, 1997. (b) Shriver, D. F.; Drezdson, M. A. The Manipulation of Air-Sensitive Compounds, 2nd ed.; Wiley: New York, 1986. (20) Kosolapoff, G. M.; Maier, L. Organic Phosphorus Compounds; Wiley: New York, 1972. (21) Lapido, F. T.; Anderson, G. K. Organometallics 1994, 13, 303306. (22) Appleton, T. G.; Bennett, M. A.; Tomkins, I. B. J. Chem. Soc., Dalton Trans. 1976, 439-446. (23) Clark, H. C.; Manzer, L. E. J. Organomet. Chem. 1973, 59, 411423. (24) Piper, T. S.; Wilkinson, G. J. Inorg. Nucl. Chem. 1956, 3, 104124. (25) Edgell, W. F.; Lyford, J. Inorg. Chem. 1970, 9, 1932-1933. (26) Ruff, J. K.; Schlientz, W. J. Inorg. Synth. 1974, 15, 84-90.

Fukuoka et al. (dppe)(CH3)Pd-MoCp(CO)3 (2). This compound was obtained as yellow-brown needles from a mixture of toluene and hexane. Yield: 38%. Mp: 122 °C dec. Anal. Calcd for C35H32O3P2MoPd: C, 54.96; H, 4.22. Found: C, 55.35; H 4.48. Λ (THF, 24 °C): 0.0069 S cm2 mol-1. IR (νCO, KBr): 1901, 1792 cm-1. 1H NMR (C6D6): δ 1.52 (dd, 3JPH ) 6.0, 7.8 Hz, 3H, CH3), 1.7-1.8 (m, 4H, dppe CH2), 4.69 (s, 5H, Cp), 7.0-7.8 (m, dppe C6H5). 31P{1H} NMR (C6D6): δ 35.4 (d, 2JPP ) 18 Hz), 56.8 (d, 2J PP ) 18 Hz). (dppe)(CH3)Pd-WCp(CO)3 (3). This compound was obtained as yellow-brown needles from toluene/hexane. Yield: 29%. Mp: 108 °C dec. Λ (THF, 24 °C): 0.0038 S cm2 mol-1. This complex was identified by spectroscopic methods. IR (ν(CO), KBr): 1893, 1785 cm-1. 1H NMR (C6D6): δ 1.69 (t, 3JPH ) 6.9 Hz, 3H, CH3), 1.6-1.7 (m, 4H, dppe CH2), 4.64 (s, 5H, Cp), 6.9-7.8 (m, 20H, dppe C6H5). 31P{1H} NMR (C6D6): δ 38.2 (d, 2JPP ) 18 Hz), 60.7 (d, 2JPP ) 18 Hz). The dinuclear ethyl complexes (dppe)(CH3CH2)Pt-MLn (MLn ) MoCp(CO)3, WCp(CO)3, and Co(CO)4) were prepared according to the procedures reported previously.7e X-ray Structure Analyses. A Rigaku four-circle diffractometer with graphite-monochromatized Mo KR radiation (0.710 69 Å) was used for data collection. The unit cells were determined by the automatic indexing of 25 centered reflections. Data were measured via ω-scans and corrected for Lorentz and polarization effects. The structures were solved and refined using the TEXSAN program system (Rigaku). Complex 4. A selected single crystal with dimensions 0.19 × 0.19 × 0.06 mm was mounted in a thin glass capillary (GLASS 0.7 mmf) under N2. Reflection data were collected at 296 K, and out of 7052 unique reflections, 904 (|Fo| > 3σ|Fo|) were observed. The structure was solved by Patterson methods and refined by a full-matrix least-squares procedure. All the non-hydrogen atoms except C3, C7, C14-C16, C19-C21, and C23-C24 were refined anisotropically. Hydrogen atoms were included in the calculation, but they were not refined. Crystallographic data for 4: C31H27O4P2CoPd, fw ) 690.83, monoclinic, P21/n (No. 14), red, a ) 10.782(8) Å, b ) 17.681(6) Å, c ) 15.950(7) Å, β ) 103.59(4)°, V ) 2955(2) Å3, T ) 296 K, Z ) 4, R (Rw) ) 0.073 (0.045), GOF ) 1.87. Complex 5. A selected single crystal with dimensions 0.50 × 0.41 × 0.16 mm was mounted in a thin glass capillary (GLASS 0.7 mmf) under N2. Reflection data were collected at 296 K, and out of 5421 unique reflections, 2709 (|Fo| > 3σ|Fo|) were observed. The structure was solved by direct methods and refined by a full-matrix least-squares procedure. All the non-hydrogen atoms were refined anisotropically. Hydrogen atoms were included in the calculation, but they were not refined. Crystallographic data for 5: C31H27O4P2CoPt, fw ) 779.52, monoclinic, P21/n (No. 14), orange, a ) 10.741(6) Å, b ) 17.716(7) Å, c ) 15.982(5) Å, β ) 103.56(3)°, V ) 2956(2) Å3, T ) 296 K, Z ) 4, R (Rw) ) 0.049 (0.038), GOF ) 1.81. Synthesis of Acyl Complexes by CO Insertion. A typical preparative procedure for (dppe)(CH3C(O))Pd-Co(CO)4 (6) is given. 4 (97.3 mg, 0.141 mmol) was dissolved in benzene under N2, and the solution was degassed in a Schlenk tube. CO (1 atm) was introduced into the Schlenk tube, and the orange solution turned to red after stirring for 16 h. The mixture was filtered, and the filtered solution was evaporated to dryness. The resulting solids were extracted with benzene, and addition of hexane to the concentrated solution of extracts gave yellow needles. The crystals were washed with cold hexane and dried under vacuum (86.4 mg, 0.120 mmol). Yield: 85%. Mp: 118 °C dec. Anal. Calcd for C32H27O5P2CoPd: C, 53.47; H, 3.79. Found: C, 53.43; H 3.88. Λ (THF, 24 °C): 0.18 S cm2 mol-1. IR (ν(CO), KBr): 2025, 1951, 1919, 1893, 1680 cm-1. 1H NMR (C6D6): δ 2.27 (s, 3H, C(O)CH3), 1.6-1.9 (m, 4H, dppe CH2), 7.0-7.6 (m, 20H, dppe C6H5). 31P{1H} NMR (C6D6): δ 24.1 (d, 2J 2 13C{1H} NMR (C D ): δ PP ) 45 Hz), 26.0 (d, JPP ) 45 Hz). 6 6 25.4 (dd, JPC ) 23, 14 Hz, dppe CH2), 27.3 (dd, JPC ) 27, 19

CO Insertion into a Pd-C Bond Hz, dppe CH2), 39.1 (dd, JPC ) 42, 21 Hz, CH3), 129-134 (m, dppe C6H5), 209-212 (m, CO), 235.0 (dd, JPC ) 115, 12 Hz, C(O)CH3). NMR Tube Reactions for CO Insertion. A typical procedure for 4 is given. CD2Cl2 (600 µL) was vacuumtransferred into an NMR sample tube (5 mmf × 180 mm) containing 4 (9.8 mg, 0.014 mmol), and 1,4-dioxane (0.5 µL) was added as an internal standard to the solution ([4] ) 0.023 M). The solution was freeze-pump-thaw degassed, and CO (1 atm, ca. 9-fold excess of 4) was introduced. The sample tube was placed in a thermostat at 24 ( 1 °C, and 1H NMR spectra were periodically measured to follow the CO insertion reaction. Reaction of 4 with dppe or PMe3. A typical procedure for PMe3 is shown. To a benzene solution of 4 (99.7 mg, 0.144 mmol) was added P(CH3)3 (0.145 mmol). Stirring at room temperature immediately gave a red oil and a yellow supernatant solution, and further stirring for 30 min yielded a red solution. The solution was evaporated to dryness, and the resulting solid was extracted with THF. After the filtered solution was again evaporated to dryness, the brown solids were washed with diethyl ether to give a white solid. The solid was recrystallized from a mixture of CH2Cl2 and Et2O to give white crystals (57.7 mg, 0.0752 mmol). Yield: 52%. Anal. Calcd for C34H36O4P3CoPd: C, 53.25; H, 4.73. Found: C, 53.19; H, 4.23. IR (νCO, KBr): 1885, 1865 cm-1. 1H NMR (CD2Cl2): δ 0.48 (q, 3JPH ) 6.4 Hz, 3H, CH3), 1.21 (dd, JPH ) 9.5, 2.3 Hz, 9H, P(CH3)3), 2.41 (m, 4H, dppe CH2), 7.0-7.6 (m, 20H, dppe C6H5). 31P{1H} NMR (C6D6): δ 12.0 (dd, 2JPP ) 387, 34 Hz, P(CH3)3), 71.6 (dd, 2JPP ) 34, 27 Hz, dppe P cis to P(CH3)3), 83.6 (dd, 2JPP ) 387, 27 Hz, dppe P trans to P(CH3)3). [Pd2(CH3)2(µ-dppe)(dppe)2][Co(CO)4]2 was obtained as pale yellow crystals from a mixture of CH2Cl2 and Et2O. Yield: 82%. This complex was identified by spectroscopic methods. IR (ν(CO), KBr): 1878 cm-1. 1H NMR (CD2Cl2): δ 0.22 (br s, 6H, CH3), 2.1-2.3 (m, 12H, dppe CH2), 6.8-7.5 (m, 60H, dppe C6H5). 31P{1H} NMR (CD2Cl2): δ 22.4 (dm, 2JPP ) 370 Hz, µ2dppe), 42.9 (dd, 2JPP ) 32, 27 Hz, µ1-dppe P trans to CH3), 59.6 (dd, 2JPP ) 370, 27 Hz, µ1-dppe P cis to CH3). CO Insertion using 13CO. 4 (56.0 mg, 0.0811 mmol) was dissolved in benzene in a Schlenk tube under N2, and the solution was degassed. A known amount of 13CO (2.22 mmol) was measured and adsorbed on dry silica gel at -196 °C using a manometer with a vacuum line. The 13CO gas was introduced into the Schlenk tube, where the CO partial pressure was roughly 1 atm. The mixture was stirred at room temperature for 6 h, and the orange solution turned to red. The mixture was evaporated to dryness, and the resulting solid (6*) was characterized by IR and NMR. Selected spectroscopic data for 6* are as follows. IR (ν(CO), KBr): 2002, 1976, 1901, 1875, 1855, 1818, 1674, 1639 cm-1. 31P{1H} NMR (C6D6): δ 24.2 (dd, 13 2J 2 2 PP ) 44, JCP ) 115 Hz, P trans to CH3 C(O), d, JPP ) 44 Hz, P trans to CH312C(O)), 25.9 (br d, 2JPP ) 44 Hz, P trans to Co). 13C{1H} NMR (C6D6): δ 209-212 (m, CO), 235.0 (dd, 2JPC ) 115, 12 Hz, CH3C(O)).

Organometallics, Vol. 20, No. 10, 2001 2075 Methods of Calculations. The B3LYP density functional theory was adopted in all the calculations.27 In determining the structures of stationary points we used the LANL2DZ basis set implemented in the Gaussian series of programs, which consists of Los Alamos effective core potentials28 for the Pd, Co, and P atoms and D95 split valence basis functions29 for the C, O, and H atoms. Polarization functions were added on the P atoms.30 This set of basis functions is called the basis set I. Using the structures thus determined, energy calculations were also carried out using the larger basis set II with d polarization functions30 on the C and O atoms and f polarization functions31 on the Pd and Co atoms, to obtain more reliable energetics. The optimized geometries of transition states (TSs) as well as equilibrium structures were characterized by normal-coordinate analysis, and the intrinsic reaction coordinates were followed to identify the reactant and product for each TS. The unscaled vibrational frequencies at the B3LYP/I level were used in the calculations of enthalpy. All the calculations were performed using the Gaussian 98 program.32

Acknowledgment. Part of the calculations were carried out at the Computer Center of the Institute for Molecular Science. This work was supported by a Grantin-Aid for Scientific Research from the Ministry of Education, Science, Sports and Culture of Japan. Supporting Information Available: Tables of crystal data and structure solution and refinement details, atomic coordinates, bond lengths and angles, and anisotropic thermal parameters for 4 and 5. This material is available free of charge via the Internet at http://pubs.acs.org. OM001037V (27) (a) Becke, A. D. Phys. Rev. A 1988, 38, 3098-3100. (b) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785-789. (c) Becke, A. D. J. Chem. Phys. 1993, 98, 5648-5652. (28) (a) Wadt, W. R.; Hay, P. J. J. Chem. Phys. 1985, 82, 284-298. (b) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 270, 299-310. (29) Dunning, T. H., Jr.; Hay, P. J. In Modern Theoretical Chemistry; Schaefer, H. F., III, Ed.; Plenum: New York, 1976; Vol. 3, pp 1-29. (30) Huzinaga, S.; Andzelm, J.; Klobukokowski, M.; Radzio-Andzelm, E.; Sakai, Y.; Tatewaki, H Gaussian Basis Sets for Molecular Calculations; Elsevier: Amsterdam, 1984. (31) Ehlers, A. W.; Bo¨hme, M.; Dapprich, S.; Gobbi, A.; Ho¨llwarth, A.; Jonas, V.; Ko¨hler, K. F.; Stegmann, R.; Veldkamp, A.; Frenking, G. Chem. Phys. Lett. 1993, 208, 111-114. (32) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.; Stratmann, R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi, M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski, J.; Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz, J. V.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P. M. W.; Johnson, B. G.; Chen, W.; Wong, M. W.; Andres, J. L.; Head-Gordon, M.; Replogle, E. S.; Pople, J. A. Gaussian 98; Gaussian, Inc.: Pittsburgh, PA, 1998.